Hyaloraphidium Curvatum Is Not a Green Alga, but a Lower Fungus; Amoebidium Parasiticum Is Not a Fungus, but a Member of the Drips

Protist, Vol. 151, 253–262, October 2000 © Urban & Fischer Verlag http://www.urbanfischer.de/journals/protist Protist

ORIGINAL PAPER

Hyaloraphidium curvatum is not a Green Alga, but a Lower Fungus; Amoebidium parasiticum is not a Fungus, but a Member of the DRIPs

Iana Ustinovaa, Lothar Krienitzb, and Volker A. R. Hussa,1 a Department of Botany, University of Erlangen-Nürnberg, D-91058 Erlangen, Germany b Institute of Freshwater Ecology and Inland Fisheries, D-16775 Neuglobsow, Germany

Submitted March 3, 2000; Accepted August 2, 2000 Monitoring Editor: Mitchell L. Sogin

The unicellular heterotrophic protist Hyaloraphidium is classified with a family of green algae, the Ankistrodesmaceae. The only species that exists in pure culture and that is available for taxonomic studies is H. curvatum. Comparative18S ribosomal RNA sequence analyses showed that H. curvatum belongs to the fungi rather than to the algae. Within the fungi, H. curvatum preferentially clustered with Chytridiomycetes. Unlike Chytridiomycetes, H. curvatum propagates by autosporulation, and the presence of flagella has never been reported. Transmission electron microscopy indicated that H. curvatum in some respects resembles Chytridiomycetes, but no elements of a flagellar apparatus were detected. The habitus of H. curvatum is unlike that of other fungi except the trichomycete Amoebidium parasiticum. The cell wall sugar composition of H. curvatum was unique, but to some extent resembled that of A. parasiticum. However, H. curvatum and A. parasiticum are not closely related to each other according to 18S rRNA sequence data. Moreover, A. parasiticum clustered with protistan animals, the Mesomycetozoa (DRIPs). Combined molecular, ultrastructural and chemical data do not allow assignment of H. curvatum to any recognized clade of fungi. This suggests that H. curvatum may represent an independent evolutionary lineage within the fungi.

Introduction

The lack of obvious morphological characters com- vent such problems and provide an independent bined with an exclusively asexual life cycle can method for drawing taxonomic conclusions. They cause considerable problems in the taxonomic dif- not only allow distinction between taxa, but also can ferentiation and identification of unicellular organ- reveal the fundamental systematic position of dubi- isms. Similarities in morphology and vegetative ous species. As a consequence, classifications propagation may be indicative of evolutionary rela- based on molecular data are sometimes not in tionships, but could also occur due to convergence, agreement with the traditional view. Examples are especially when the same ecological niche is not rare, where species previously assigned to fungi shared. In contrast, molecular data usually circum- turned out to be plants or protists and vice versa (e. g. Herr et al. 1999; Huss and Sogin 1990; Ragan et al. 1996). 1 Corresponding author; fax 49-9131-8528751 The genus Hyaloraphidium Pascher et Korshikov e-mail [email protected] was first established in 1931 for some autosporine

1434-4610/00/151/03-253 $ 15.00/0 254 I. Ustinova et al. and colourless green alga-like organisms (Ko- pods, usually as endocommensals of the gut. Origi- rschikoff 1931). Based on similarities in morphology nally, the Trichomycetes consisted of four orders: and mode of reproduction, a close relationship to Amoebidiales, Eccrinales, Asellariales, and Harpel- the genus Ankistrodesmus Corda (Ankistrodes- lales (Duboscq et al. 1948, Manier 1955). Today, maceae) was assumed. H. contortum Pasch. et Harpellales and Asellariales are proposed to belong Korsh. with thin needle-like screwed cells similar to to the Zygomycetes, although only some members Ankistrodesmus spiralis was described as the type of Harpellales produce zygospores (Lichtwardt species, but H. curvatum Korsh., used in our study, 1986; O'Donnell et al. 1998). Combining the remain- is the only species that has been isolated into pure ing orders, Amoebidiales and Eccrinales, in Tri- culture until now. It has thicker, shorter, and more or chomycetes is thought to be artificial (Lichtwardt less twisted cells compared to H. contortum, and 1986). This view is based on differences in morphol- was isolated from a strain of the chlorophyte Hydro- ogy, ecology and development and is further sup- dictyon reticulatum (Pringsheim 1963). ported by differences in their cell wall chemistry The relationship of Hyaloraphidium to the family (Lichtwardt 1986; Whisler 1963). Ankistrodesmaceae (syn. Selenastraceae) has been A. parasiticum was isolated into pure culture by the subject of many discussions. Hindák (1977) con- Whisler (1960, 1962). Its development, cell wall sidered some species of Hyaloraphidium to be chemistry, and ultrastructure were extensively stud- colourless varieties of Monoraphidium Komárková- ied (Trotter and Whisler 1965; Whisler1962; Whisler Legnerová (Ankistrodesmaceae), while Kiss (1984) 1968; Whisler and Fuller 1968). Whisler (1963) questioned its affiliation to any alga. One species, H. doubted the fungal origin of the Amoebidiales and moinae Korsh., has been proposed to be synony- suggested their position to be “somewhere between mous to Amoebidium parasiticum Cienk. (Amoebidi- fungi and protozoa”. Later, based on 5S rRNA se- ales), which is classified with the fungal class Tri- quences, A. parasiticum was shown to be more chomycetes (Hindák 1977; Komárek and Fott 1983). closely related to the protozoa than to the Zygomy- Colourless algae of the genus Gloxidium Korsh., de- cota (Walker 1984). However, the data were not suf- scribed as closely related to Hyaloraphidium and ficient to allow more detailed phylogenetic conclu- Ankistrodesmus (Korschikoff 1931), were also pro- sions. posed to belong to the Amoebidiales (Hindák 1977; Here we present results of phylogenetic analyses Pringsheim 1963). of 18S rRNA gene sequences, electron microscopy, Amoebidium parasiticum is mentioned several and cell wall chemical analyses to discuss the rela- times in context with Hyaloraphidium species. It re- tionships between Hyaloraphidium curvatum and sembles Hyaloraphidium in many features such as (i) Ankistrodesmaceae, and between H. curvatum and thallus morphology; (ii) sporangia with regularly ar- Amoebidium parasiticum. ranged sporangiospores that look like autospores in a mother cell of Hyaloraphidium. Moreover, in pure culture A. parasiticum does not undergo an amoe- Results boid stage in its life cycle, but reproduces only by nonmotile sporangiospores (Whisler 1962). How- Phylogenetic Position of Hyaloraphidium ever, the systematic position of Amoebidium itself is curvatum not clear. The genus Amoebidium with A. parasiticum as its type species was established by The amplification of the 18S rRNA gene of H. curva- Cienkowski in 1861 for the exocommensal epizoon tum resulted in a product of 1796 base pairs without found on a variety of aquatic arthropods. It has a introns. Its sequence was determined and com- colourless cigar-shaped thallus and a basal hold- pared with the EMBL database. Unexpectedly, the fast. Reproduction occurs either by nonmotile spo- most similar sequences were found in fungi and not rangiospores or by amoebae. The production of in the green algae. The four highest similarities were amoebae within the sporangium was thought to be obtained with Spizellomyces acuminatus (Chytrid- a specific feature of this organism. Cienkowski iomycota; 92.1%), Basidiobolus ranarum (Zygomy- (1861) assigned A. parasiticum to the fungal order cota; 90.7%), the Lasioderma serricorne yeast-like Chytridiales, but others classified it with the Proto- symbiont, (Ascomycota; 90.1%), and Rhizoctonia zoa, usually with Sporozoa or Myxomycophyta (see solani (Basidiomycota; 90.1%). In order to elucidate Whisler 1962). Later, Amoebidium was assigned to the phylogenetic position of H. curvatum, we con- the class Trichomycetes (Zygomycota) (Duboscq et structed phylogenetic trees with the 18S rRNA gene al. 1948). This class combines unique fungi that are sequences of representatives of all major fungal found in permanent association with various arthro- groups (Fig. 1A). In all analyses, H. curvatum was Hyaloraphidium and Amoebidium Phylogeny 255

Figure 1. (A) Neighbor-joining tree inferred from 18S rRNA gene sequences of Hyaloraphidium curvatum, Amoe- bidium parasiticum, and 22 other eukaryotes. Numbers above and below the forks give the bootstrap support (100 replicates) for neighbor-joining (NJ), maximum parsimony (MP), and maximum likelihood (ML) analyses. Only bootstrap values above 50% are shown. Asterisks indicate that the topology obtained by the MP and ML analyses was different, but statistically unsupported. (B) Maximum likelihood tree inferred from 18S rRNA sequences of A. parasiticum and members of the DRIPs (Mesomycetozoa). NJ, MP, and ML tree topologies are congruent. Numbers above and below forks as in (A), but for 1000, 500, and 100 replicates in the NJ, MP, and ML analyses respectively. 256 I. Ustinova et al. Hyaloraphidium and Amoebidium Phylogeny 257 preferentially grouped within a clade of fungi that in- genesis. Mitochondria are elongated and some- cluded Chytridiomycetes and some Zygomycetes. times even thread-like (Fig. 2C, 2G). Numerous In spite of its tendency to group with Chytridio- larger organelles, which are covered with only one mycetes, the exact phylogenetic position of H. cur- membrane and have a homogeneous content, may vatum within the fungi could not be resolved, except be microbodies. They are sometimes associated that a relationship with higher fungi, Ascomycetes with lipid globules, mitochondria and cisternae of and Basidiomycetes, was excluded. the endoplasmic reticulum (Fig. 2H). To some extent, this complex resembles the microbody lipid Transmission Electron Microscopy of globule complex (MLC) of zoospores of chytrid- H. curvatum iomycetes (Barr 1990; Berger et al. 1998). The abun- dant vacuoles often include osmiophilic dense ma- The morphology and reproduction of H. curvatum terial or membranous structures. In some cells un- do not resemble that of any described chytrid- usual structures were observed that are not en- iomycete or zygomycete. Typical cells are more or closed by any membrane and have a regularly less curved with rounded ends and without any rhi- striped or chequered pattern (Fig. 2G, 2I). These zoids or protrusions (Fig. 2A). “Autospores”, similar structures resemble the rumposomes of chytrid- to those of coccal green algae, are formed within the iomycetes in a face-on view (Barr 1990). However, mother cell and released by rupture of its wall (Fig. they are not aggregated with lipid globules, micro- 2B). The number of spores per cell is 4-8, but some- bodies, or vacuoles, and a view that would be typi- times is irregular. The ultrastructure of spores does cal for a rumposome in cross-section was never not differ from that of mature cells. Nuclei are chro- found. On the other hand, structures with such a matin-rich and often have a bulge filled with an regular pattern are characteristic of phospholipids amorphous material associated with the nuclear en- which are often incorporated in so-called “dense velope (Fig. 2A, 2C). This structure could be inter- bodies”. Dense bodies are common in many fungi, preted as a nucleolus or an intranuclear nucleus-as- but they are surrounded by a single membrane sociated organelle (NAO)-like structure of unknown (Beckett et al. 1974). Sections through numerous function. However, identification of this structure re- mature cells and spores, and even serial cross-sec- quires additional cytochemical analyses. Spores tions through several cells gave no evidence that contain only one nucleus, but mature cells with up to centrioles, basal bodies, or flagella are present in H. four nuclei were observed (Fig. 2C). Fig. 2D shows a curvatum. nucleus in the metaphase of mitotic division. As chromatin aligns along the metaphase plate, the Cell Wall Sugar Composition of H. curvatum “nucleolus” or “NAO” is not visible any more, and the spindle microtubules are seen to diverge from The cell wall sugar composition may be an impor- pocket-like regions at the spindle poles of the intact tant feature at certain taxonomic levels for the sys- nuclear envelope. This kind of mitotic division re- tematics of fungi. In H. curvatum, the fibres of the sembles that shown in a chytridiomycete, Catenaria TFA-resistant, alkali-insoluble rigid cell wall fraction anguillulae, except that centrioles are not present consisted of equal amounts of galactose and man- near the “pockets” (Beckett et al. 1974). A ribosomal nose, while the TFA-hydrolyzable matrix was com- core area around the nucleus was never observed in posed of three-quarters galactose and one-quarter mature cells or in autospores. mannose. Glucose, glucosamine or chitin were not Unlike most fungi, H. curvatum possesses found in fibre or matrix fractions. The crystallinity of stacked Golgi dictyosomes (Fig. 2E, 2F) which ac- the cell wall was low, as it did not brighten in a dark tively take part in the synthesis of new cell walls by field under polarized light. This kind of cell wall is producing a large number of vesicles during sporo- very different from that of most other fungi. How-

Figure 2. Electron micrographs of Hyaloraphidium curvatum. (A) Typical cell appearance. The arrow shows the bulge of the nucleus filled with an amorphous material, either nucleolus or NAO. (B) “Autospores” are released from the mother cell. No ribosomal core area is formed around nuclei. (C) Cell with two nuclei. The filled bulge of one nucleus is indicated by the arrow. (D) Nucleus in the metaphase of mitotic division. The nucleus envelope is intact. The microtubuli of the mitotic spindle are visible. (E–F) Dictyosomes with stacked Golgi cisternae. (G) Cross-section of the cell showing the thread-like mitochondria and a structure with “chequered” pattern (arrow). (H) MLC – like aggregation. (I) “Chequered” structure which resembles a rumposome of Chytridiomycetes in face-on view. n – nucleus; m – mitochondrium; mb – microbody; l – lipid globule, er – endoplasmatic reticulum. Bars indicate 1 µm (A–C, G, H) or 0.1 µm (D–F, I). 258 I. Ustinova et al. ever, galactan along with galactosamine are known concordance with an analysis based on combined as main constituents in cell walls of Amoebidium 18S, 5.8S, and 28S rRNA data (Van der Auwera and parasiticum (Trotter and Whisler 1965). For this and De Wachter 1996). Other Chytridiomycetes with other similarities between H. curvatum and A. para- available 18S rRNA sequences belong to the order siticum mentioned before, the phylogenetic relation- Neocallimasticales and cluster with Neocallimastix ship between the two organisms was determined. joyonii (not shown). The position of the flagella-lack- ing zygomycete Basidiobolus ranarum within the Phylogenetic Position of Amoebidium Chytridiomycetes rather than the Zygomycetes has parasiticum previously been discussed by others (Jensen et al. 1998; Nagahama et al. 1995). In general, the phy- The 18S rRNA gene sequence of A. parasiticum is logeny of lower fungi is poorly resolved and a mono- 1797 base pairs long and contains no introns. The phyletic origin of Zygomycetes or Chytridiomycetes highest sequence identity (95.3%) was found with is not statistically supported by ours or previous the fish parasite Ichthyophonus hoferi which be- analyses based on 18S rRNA data (Jensen et al. longs to a group of protists that evolved near the an- 1998; Nagahama et al. 1995). imal-fungal divergence (DRIPs) (Ragan et al. 1996), The suggested relationship of H. curvatum with recently named Mesomycetozoa (Herr et al. 1999). the clade of Chytridiomycetes is rather surprising. The phylogenetic tree in Fig. 1B clearly shows that Traditionally, Chytridiomycetes are defined as fungi A. parasiticum is a member of the DRIPs and most that always have flagellate stages in their life cycle. closely related to I. hoferi. This phylogeny is highly However, a flagellate stage has never been de- supported with bootstrap values of 100%. The scribed for H. curvatum. As some Zygomycetes also DRIPs constitute a well-resolved clade which cluster with Chytridiomycetes, H. curvatum might groups with animals, and not with fungi (Fig. 1A). It is represent a transition form from flagellated to non- split into two distinctive clusters. The first includes I. flagellated fungi. Alternatively, flagella might have hoferi and A. parasiticum along with Psorospermium been lost during the evolution of H. curvatum, or haeckelii, Anurofeca richardsi, and Sphaerosoma may be produced only at a distinctive stage of its life arcticum. The second contains Dermocystidium, cycle. In this case, the potential existence of flagella Rhinosporidium seeberi, and the “rosette agent”. should be recognised by the presence of basal bod- In contrast to some similarities in morphology and ies or centrioles. In contrast to Chytridiomycetes, cell wall chemistry, no evidence was found for a Zygomycetes and higher fungi do not possess cen- close genetic relationship of Hyaloraphidium curva- trioles but nucleus-associated organelles (NAOs). tum and Amoebidium parasiticum. Thus, the presence of either centrioles or NAOs could help to resolve the systematic position of H. curvatum at least at the class level. Discussion The results of transmission electron microscopy of H. curvatum do not elucidate its relationship The earlier identification of Hyaloraphidium curva- within fungi. On one hand, the presence of the tum as a colorless green alga was based entirely on stacked Golgi apparatus (Fig. 2E, 2F) typical for its similarity to species of the green algal family Chytridiomycetes but not for other fungi except Tri- Ankistrodesmaceae in vegetative reproduction by chomycetes (Beckett et al. 1974; Weber 1993), and 4–8 “autospores” and in cell morphology as ob- a kind of microbody-lipid globule complex (MLC) served under the light microscope. Our analyses of (Fig. 2H) supports the preferential clustering of H. 18S rRNA gene sequences demonstrate that H. cur- curvatum with Chytridiomycetes. Moreover, if the vatum does not belong to the Ankistrodesmaceae, “chequered” organelle (Fig. 2G, 2I) is a rumposome, but to lower fungi (Fig. 1A). However, the precise H. curvatum could be classified with either the systematic position of H. curvatum could not be re- Chytridiales or the Monoblepharidales, as rumpo- solved despite its preference to cluster with the somes are characteristic for these orders (Barr Chytridiomycetes. Within our taxon sampling, H. 1990; Weber 1993). On the other hand, thallus mor- curvatum constitutes the most basal branch within phology and development of H. curvatum is different Chytridiomycetes except for Blastocladiales, but the from Chytridiomycetes. It does not possess a flagel- statistical support for this topology is low. Blastocla- lar apparatus characteristic of Chytridiomycetes or diales, represented by Allomyces macrogynus and a ribosomal core area which is present in all chytrid- Blastocladiella emersonii, group together indepen- iomycetes except Spizellomycetales (Barr 1990; dent of other Chytridiomycetes, and appear here to Weber 1993). Moreover, H. curvatum is not a para- represent the earliest lineage of fungal radiation in site (Korschikoff 1931; Pringsheim 1963), and its cell Hyaloraphidium and Amoebidium Phylogeny 259 wall chemistry is different from both Chytrid- 1A) suggesting that these similarities occurred due iomycetes and Zygomycetes, as it is composed of to convergence. No 18S rRNA sequences of the equal amounts of galactose and mannose. The cell second order of Trichomycetes, the Eccrinales, are wall of Chytridiomycetes mainly consists of chitin available so far. However, a relationship of H. curva- and ß-glucan, while the major cell wall components tum with Eccrinales is unlikely on the basis of differ- of Zygomycetes are chitosan and chitin (Bartnicki- ences in their habits and morphology. Garcia 1968). However, galactose along with poly- Despite obvious morphological differences, the galactosamin as major cell wall constituents were ultrastructure of H. curvatum in many aspects re- found in Amoebidium parasiticum (Trichomycetes) sembles that of Chytridiomycetes, in agreement (Trotter and Whisler 1965). with the molecular phylogeny. It is commonly be- Where is the place of H. curvatum within fungi? A lieved that Chytridiomycetes are the most primitive summary of “Pros” and “Cons” for the classification fungal group and as such still bear flagella. If the of H. curvatum with established groups of lower placement of H. curvatum within Chytridiomycetes fungi is given in Table 1. With the available data it is reflects a natural relationship than the complete loss impossible to assign H. curvatum to any of these of flagella during the evolution of H. curvatum has to groups. It is obvious from molecular data, cell wall be assumed. On the other hand, H. curvatum may chemistry and electron microscopy that H. curvatum represent a link between Chytridiomycetes and Zy- does not belong to Ascomycetes or Basidio- gomycetes. However, as the cell wall composition of mycetes. Neither could it be a zygomycete, as it has H. curvatum is so different, it seems appropriate to a stacked Golgi apparatus and its cell wall does not regard it as an independent lineage of fungal evolu- consist of chitin and chitosan. tion. Better resolution of its position within fungi will On first view, H. curvatum most closely resembles require further analyses of more molecular data. The the “trichomycete” A. parasiticum. This resem- question of systematic assignment of other species blance is based on similarities in cell wall chemistry, of Hyaloraphidium including the type species H. thallus morphology, and the observation that A. par- contortum is still open. asiticum restricts itself to a sporangiospore cycle in The results of our 18S rRNA analyses clearly pure culture (Whisler 1962), whereby unflagellated show that A. parasiticum is unrelated not only to H. spores are arranged in a sporangium in a manner curvatum but also to other fungi (Fig. 1A). Instead, it similar to that of H. curvatum. However, a relation- belongs to a certain clade of protistan parasites of ship between H. curvatum and Amoebidiales was fishes and crustaceans that was provisionally refuted by the 18S rRNA sequence analyses (Fig. named “DRIPs” after the initials of known members

Table 1. Arguments for and against assigning Hyaloraphidium curvatum to established classes of lower fungi.

Pro Contra

Chytridiomycetes: Chytridiomycetes: • Phylogenetic position • Reproduction (via non-motile “autospores”) and • Golgi apparatus with stacked cisternae ecology are different from those of Chytridiomycetes • Microbody-lipid globule complex • Absence of flagella and flagellar apparatus • Rumposome-like structure • Absence of ribosomal core area in spores • Cell wall composition

Zygomycetes: Zygomycetes: • Phylogenetic position • Thallus morphology (unicellular, fusiform, curved, • Absence of flagella and flagellar apparatus without rhizoids) and mode of propagation • Golgi apparatus with stacked cisternae • Cell wall composition

Trichomycetes: Trichomycetes: • Golgi apparatus with stacked cisternae • Phylogenetically not related to Amoebidium • Thallus morphology and mode of propagation parasiticum; molecular data for Eccrinales are not (similar to Amoebidiales) available • Absence of flagella and flagellar apparatus • Cell wall composition is different from Eccrinales • Cell wall composition (similar to Amoebidiales) • Thallus morphology, mode of propagation and ecology are different from Eccrinales 260 I. Ustinova et al. at that time: Dermocystidium, “rosette agent”, Pa. DNA was purified by a standard phenol-chloro- Ichthyophonus, and Psorospermium (Ragan et al. form procedure and precipitated with ethanol. 18S 1996). Recently, the name Mesomycetozoa was rRNA genes were amplified by PCR and directly se- proposed (Herr et al. 1999), reflecting their relation quenced either manually (H. curvatum) with the T7 to both fungi and animals, as they form the most Sequenase PCR Product Sequencing Kit (Amer- basal branch within animals just after the point of sham) or with the Big Dye Terminator Cycle se- fungal-animal divergence (Ragan et al. 1996). quencing kit (PE Applied Biosystems) in the ABI Among the DRIPs, Ichtyophonus hoferi shares the Prism 310 Genetic Analyser of Perkin-Elmer (A. par- highest 18S rRNA similarity with A. parasiticum (Fig. asiticum). PCR and sequencing primers were the 1B). Some morphological and developmental fea- same as described previously (Huss et al. 1999). tures support this relation as well. Both I. hoferi and 18S rRNA sequences of H. curvatum and A. para- A. parasiticum have amoeboid but no flagellate siticum were deposited in GenBank under the ac- stages in their life cycle, both possess NAOs, and cession numbers Y17504 and Y19155 respectively. both do not contain chitin in their cell wall. However, Sequence data were compared with sequences mitochondria of I. hoferi do not have flat but the pre- in the EMBL database by the use of the FASTA3 al- sumptively more primitive tubulovesiculate cristae. gorithm (Pearson 1990). A subalignment of the most Although the shape of mitochondrial cristae is in similar sequences found and of representatives of general a good indicator for the phylogenetic posi- all major fungal groups was obtained from the RDP tion, exceptions are known (see Ragan et al. 1996). database (http://www.cme.msu.edu/rdp/html/index. The tubulovesiculate form of cristae in I. hoferi has html) (Maidak et al. 1999). Sequences not included been interpreted as a reversion from the flat form in the RDP alignment were added manually consid- typical for animals and fungi (Ragan et al. 1996). The ering the conserved secondary structure (Huss and fact that Amoebidiales are closely related to I. hoferi Sogin 1990; Neefs and DeWachter 1990). The fol- supports this hypothesis, as they possess flat mito- lowing reference sequences were used in the chondrial cristae (Dang and Lichtwardt 1979; phylogenetic analyses (GenBank accession num- Whisler and Fuller 1968). The number of genera as- bers are given in parentheses): Tripedalia cysto- signed to the Mesomycetozoa continues to grow phora (L10829), Anemonia sulcata (X53498), Ich- (Baker et al. 1999; Herr et al. 1999; Joestensen et thyophonus hoferi (U25637), Anurofeca richardsi al., unpublished). According to the phylogenetic tree (AF070445), Sphaerosoma arcticum (Y16260), in Fig. 1B, Mesomycetozoa are split into two distinc- Psorospermium haeckelii (U33180), “rosette agent” tive groups, one of which contains all the species (choanoflagellate-like sp.) (L29455), Rhinosporidium with an amoeboid life stage. seeberi (AF118851), Dermocystidium salmonis Although a relationship of A. parasiticum to proto- (U21337), Dermocystidium sp. (U21336), Blasto- zoa rather than to fungi was suggested by earlier cladiella emersonii (M54937), Allomyces macrogy- studies (Walker 1984), its precise assignment to any nus (U23936), Sporidiobolus johnsonii (L22261), particular group of protists could not be achieved. Ustilago maydis (X62396), Saccharomyces cerevi- This is now possible with our molecular analyses. siae (Z75578), Aspergillus fumigatus (AB008401), Moreover, our data suggest complete revision of the Lasioderma serricorne yeast-like symbiont (D49656), systematics of the whole order Amoebidiales on a Glomus intraradices (AF004681), Acaulospora spi- molecular basis. nosa (Z14004), Endogone pisiformis (X58724), Mortie-rella polycephala (X89436), Chytridium con- fervae (M59758), Basidiobolus ranarum (D29946), Methods Neocallimastix joyonii (M62705), Spizellomyces acuminatus (M59759); Chlorella vulgaris (X13688), Hyaloraphidium curvatum SAG 235-1 was obtained and Ankistrodesmus stipitatus (X56100). The 18S from the Collection of Algae, Göttingen, Germany, rRNA gene sequences of the oomycete Achlya bi- and Amoebidium parasiticum ATCC 32708, from the sexualis (M32705) and of the slime mould Dictyo- American Type Culture Collection, Manassas, USA. stelium discoideum (X00134) were used as out- H. curvatum was cultured in the recommended groups. Highly variable regions that could not be Polytoma medium (Schlösser 1994), and A. para- aligned unambiguously for all sequences were ex- siticum in an ATCC Tryptone-Glucose medium. Cells cluded, resulting in a total of 1572 (fungi) and of were grown at room temperature for 4–5 days with 1647 positions (DRIPs) that were used for the analy- either aeration or shaking to prevent sedimentation. ses. Phylogenetic trees were inferred by the neigh- Approximately 2 g of cells were ground in a cell mill bor-joining (NJ), the maximum parsimony (MP), and or treated in a French Pressure Cell (Aminco) at 106 the maximum likelihood (ML) methods using the Hyaloraphidium and Amoebidium Phylogeny 261

PHYLIP program package version 3.572c (Felsen- Acknowledgements stein 1995) for NJ and MP analyses of fungal 18S rRNA sequences (Fig. 1A), and the PAUP program We thank K. P. Gaffal and G. Friedrichs for super- package 4.0 b3a (Swofford 2000) for all other analy- vision and help in transmission electron microscopy, ses. For NJ, distance matrices were corrected with and G. Steingräber for excellent technical assis- the two-parameter method of Kimura (1980), and the tance. H. Takeda (Hokkaido University School of distances subsequently converted into phylogenetic Medicine, Sapporo, Japan) kindly performed the cell trees using the neighbor-joining method of Saitou wall sugar composition analyses. We also thank E. and Nei (1987). The addition of sequences was jum- Schnepf, J. Longcore, B. F. Lang, and H. C. Whisler bled, and a transition/transversion ratio of 2.0 was for helpful discussions and advice. This work was selected. The same ratio was used also for MP anal- supported by the Deutsche Forschungsgemein- yses. The search was fully heuristic with random ad- schaft Grant Hu 410/6 to V.A.R.H. dition of taxa repeated five times. The robustness of the obtained tree topologies was estimated by the bootstrap method (Felsenstein 1985). 100 bootstrap References resamplings were performed for each method with the PHYLIP program and 1000 resamplings for NJ Baker GC, Beebee TJ, Ragan MA (1999) Prototheca with PAUP. For MP analyses with PAUP, characters richardsi, a pathogen of anuran larvae, is related to a were weighted with rescaled consistency index and clade of protistan parasites near the animal-fungal divergence. Microbiology UK 145: 1777–1784 then used as input for a bootstrap analysis (500 replications). Search settings were the same as de- Barr DJS (1990) Phylum Chytridiomycota. In Margulis scribed above. Tree-bisection-reconnection (TBR) L, Corliss JO, Melkonian M, Chapman DJ (eds) Hand- was chosen as a branch-swapping algorithm. For book of Protoctista. Jones and Bartlett Publ, Boston, pp 454–466 ML analyses, the Hasegawa-Kishino-Yano model of evolution and a transition/transversion ratio of 2.0 Bartnicki-Garcia S (1968) Cell wall chemistry, morpho- were selected. Settings for the full heuristic search genesis, and taxonomy of fungi. Annu Rev Microbiol 22: were the same as for MP analyses. A bootstrap anal- 87–105 ysis was performed with 100 replications. Beckett A, Heath IB, McLaughlin DJ (1974) An Atlas Preparation of H. curvatum for transmission elec- of Fungal Ultrastructure. Longman, London tron microscopy was done as described by Gaffal Berger L, Speare R, Daszak P, Green DE, Cunning- (1987) and by Krienitz et al. (1983). After removing ham AA, Goggin CL, Slocombe R, Ragan MA, Hyatt the culture medium with 50 mM collidine buffer, cells AD, McDonald KR, Hines HB, Lips KR, Marantelli G, were fixed in collidine-buffered 2% glutaraldehyde Parkes H (1998) Chytridiomycosis causes amphibian containing 5 mM CaCl for 1 h at room temperature mortality associated with population declines in the rain 2 forests of Australia and Central America. Proc Natl and postfixed in 2% OsO and O.8% K Fe (CN) for 4 3 3 6 Acad Sci USA 95: 9031–9036 1 h at room temperature. Cells were then rinsed 3 Cienkowski L (1861) Ueber parasitische Schläuche auf times with H2O and treated for 2 h with 2% uranyl acetate. After standard dehydration, embedding, Crustaceen. Botan Zeitung 19: 167–174 and sectioning, samples were stained with lead cit- Dang S, Lichtwardt RW (1979) Fine structure of rate. Observations were made under a Zeiss EM-10 Paramoebidium (Trichomycetes) and a new species electron microscope. with viruslike particles. Am J Bot 66: 1093–1104 The rigid cell wall fraction of H. curvatum cells Duboscq O, Leger L, Tuzet O (1948) Contribution a la was isolated and analysed as described by Takeda connaissance des Eccrinides: les Trichomycetes. Arch (1991) and Krienitz et al. (1999). Cell walls of H. cur- Zool Exp Gen 86: 29–144 vatum were first refined from soluble material, Felsenstein J (1985) Confidence limits on phylogenies: treated with boiling ethanol and a boiling ether/ace- an approach using the bootstrap. Evolution 39: 783–791 tone mixture (v/v), and then successively hydrolysed Felsenstein J (1995) PHYLIP (Phylogeny Inference with trifluoracetic (TFA), sulfuric, and hydrochloric Package) Version 3.572c. Department of Genetics, Uni- acid. The composition of neutral- and amino sugars versity of Washington, Seattle, USA was determined by gas-liquid chromotography and Gaffal KP (1987) Mitosis-specific oscillations of mito- by high-voltage paper electrophoresis respectively. chondrial morphology in Chlamydomonas reinhardtii. Neutral sugars were analysed by TFA and sulfuric Endocyt Cell Res 4: 41–62 acid hydrolyses, and amino sugars by HCl hydroly- Herr RA, Ajello L, Taylor JW, Arseculeratne SN, sis. Anisotropy of isolated cell walls was examined Mendoza L (1999) Phylogenetic analysis of Rhino- under a polarizing microscope. sporidium seeberi's 18S small-subunit ribosomal DNA 262 I. Ustinova et al. groups this pathogen among members of the protoctis- Neefs J-M, DeWachter R (1990) A proposal for the tan Mesomycetozoa clade. J Clin Microbiol 37: secondary structure of a variable area of eukaryotic 2750–2754 small ribosomal subunit RNA involving the existence of Hindák F (1977) Studies on the chlorococcal algae pseudoknot. Nucleic Acids Res 18: 5695–5704 (Chlorophyceae).I. Biol Pr 23: 1–190 O'Donnell K, Cigelnik E, Benny GL (1998) Phyloge- Huss VAR, Sogin ML (1990) Phylogenetic position of netic relationships among the Harpellales and Kickxel- some Chlorella species based upon complete small- lales. Mycologia 90: 624–639 subunit ribosomal RNA sequences. J Mol Evol 31: Pearson WR (1990) Rapid and sensitive sequence 432–442 comparison with FASTP and FASTA. Methods Enzymol Huss VAR, Frank C, Hartmann EC, Hirmer M, 183: 63–98 Kloboucek A, Seidel BM, Wenzeler P, Kessler E Pringsheim EG (1963) Farblose Algen. G. Fischer-Ver- (1999) Biochemical taxonomy and molecular phylogeny lag, Stuttgart of the genus Chlorella sensu lato (Chlorophyta). J Phy- col 35: 587–598 Ragan MA, Goggin CL, Cawthorn RJ, Cerenius L, Jamieson AV, Plourde SM, Rand TG, Soderhall K, Jensen AB, Gargas A, Eilenberg J, Rosendahl S Gutell RR (1996) A novel clade of protistan parasites (1998) Relationships of the insect-pathogenic order En- near the animal-fungal divergence. Proc Natl Acad Sci tomophthorales (Zygomycota, Fungi) based on phylo- USA 93: 11907–11912 genetic analyses of nuclear small subunit ribosomal DNA sequences (SSU rDNA). Fungal Genet Biol 24: Saitou N, Nei M (1987) The neighbor-joining method: a 325–334 new method for reconstructing phylogenetic trees. Mol Biol Evol 4: 406–425 Kimura M (1980) A simple method for estimating evolutionary rates of base substitutions through comparative Schlösser UG (1994) SAG – Sammlung von Algenkul- studies of nucleotide sequences. J Mol Evol 16: 111–120 turen at the University of Göttingen. Catalogue of strains 1994. Bot Acta 107: 113–186 Kiss KT (1984) Special problems in studying phytoplankton associations (occurrence and ecological eval- Swofford DL (2000) PAUP*. Phylogenetic Analysis uation of Hyaloraphidium contortum). Acta Bot Hung Using Parsimony (*and other methods). Sinauer Associ- 30: 269–276 ates, Sunderland, Massachusetts Komárek J, Fott B (1983) Das Phytoplankton des Takeda H (1991) Sugar composition of the cell wall and Suesswassers 7,1. Chlorophyceae (Gruenalgen). Ord- the taxonomy of Chlorella (Chlorophyceae). J Phycol nung: Chlorococcales. Schweizerbart, Stuttgart 27: 224–232 Korschikoff AA (1931) Notizen über einige neue apo- Trotter MJ, Whisler HC (1965) Chemical composition chlorotische Algen. Arch Protistenkd 74: 249–258 of the cell wall of Amoebidium parasiticum. Can J Bot 43: 869–876 Krienitz L, Klein G, Heynig H, Böhm H (1983) Mor- phologie und Ultrastruktur einiger Arten der Gattung Van der Auwera G, De Wachter R (1996) Large-sub- Monoraphidium (Chlorellales) I. Monoraphidium griffi- unit rRNA sequence of the chytridiomycete Blasto- thii, M. neglectum und M. tortile. Arch Hydrobiol Suppl cladiella emersonii, and implications for the evolution of Algol Stud 33: 401–417 zoosporic fungi. J Mol Evol 43: 476–483 Krienitz L, Takeda H, Hepperle D (1999) Ultrastruc- Walker WF (1984) 5S ribosomal RNA sequences from ture, cell wall composition, and phylogenetic position of Zygomycotina and evolutionary implications. System Pseudodictyosphaerium jurisii (Chlorococcales, Chloro- Appl Microbiol 5: 448–456 phyta) including a comparison with other picoplank- Weber H (1993) Allgemeine Mykologie. G. Fischer-Ver- tonic green algae. Phycologia 38: 100–107 lag, Jena, Stuttgart Lichtwardt RW (1986) The Trichomycetes, Fungal As- Whisler HC (1960) Pure culture of the Trichomycete, sociates of Arthropods. Springer-Verlag, New York Amoebidium parasiticum. Nature 186: 732–733 Maidak BL, Cole JR, Parker CT Jr, Garrity GM, Lar- Whisler HC (1962) Culture and nutrition of Amoebidium sen N, Li B, Lilburn TG, McCaughey MJ, Olsen GJ, parasiticum. Am J Bot 49: 193–199 Overbeek R, Pramanik S, Schmidt TM, Tiedje JM, Woese CR (1999) A new version of the RDP (Ribosomal Whisler HC (1963) Observations on some new and Database Project). Nucleic Acids Res 27: 171–173 unusual enterophilous Phycomycetes. Can J Bot 41: 887–900 Manier JF (1955) Classification et nomenclature des Trichomycetes. Ann Sci Nat Zool 17: 395–397 Whisler HC (1968) Developmental control of Amoebid- Nagahama T, Sato H, Shimazu M., Sugiyama J (1995) ium parasiticum. Devel Biol 17: 562–570 Phylogenetic divergence of the Entomophthoralean Whisler HC, Fuller MS (1968) Preliminary observations fungi: evidence from nuclear 18S ribosomal RNA gene on the holdfast of Amoebidium parasiticum. Mycologia sequence. Mycologia 87: 203–209 60: 1068–1079